METHODS

In this study we enrolled 30 healthy medical students (age 22±1 yrs; height 1.69±1 m; weight 59.6±2.1 Kg) of either gender (male=10, female=20) through Campus advertising. They were studied twice: the first time (stress day) 30-60 min before an university examination and a second time (control day) three months afterwards, in a period away from study routines. In both instances, after subjects had filled out standardized questionnaires, we took a saliva sample (2 ml), in order to determine the salivary free cortisol and cytokines (IL 1, IL 2, TNF a) levels and subsequently we recorded non invasively ECG, the respiratory and the arterial pressure signals, in order to assess the autonomic control of the circulation.

All subjects were non smokers and free of any diseases, as determined by routine history and physical examination. Subjects were instructed to avoid alcohol and caffeinated beverages for the 12 hours preceding the study, to avoid heavy physical activity the day before, and to come to the laboratory after a light breakfast, between 9.30 a.m. and 11:30 a.m. The protocol of the study was approved by our Institution Review board.

Assessment of overall stress level: the overall stress level was gauged by a battery of non invasive tests: Endocrine involvement was assessed in all 30 subjects, by measuring by radioimmunoassay (DPC, Los Angeles, California, USA) free salivary cortisol level that reflects the concentration of free hormone in plasma(20). Samples both in the stress and control day were obtained at approximately the same hour (10:30AM ±1) to account for circadian variation (21). Immunological involvement was estimated in a random subgroup of 14 subjects, whose salivary sample was technically adequate, by assessing also the salivary levels of selected cytokines (IL 1, IL 2, TNF a) with commercially available ELISA assays (R&D Systems, Minneapolis, MN) following the procedures suggested by the manufactures. Prolonged exposure to stressors is signaled by alterations in cytokine profile, such as increasing IL1 and TNFa, and reducing IL2 (22).

Psychological involvement was gauged by a battery of questionnaires, providing self-rated scales, that, in line with previous (23) and more recent studies (24), focuses on appraisal, coping and health. In brief,

-The appraisal of stress, tiredness or activation were assessed by a global scoring index (0-30)

-Coping was assessed by a graphic questionnaire (score 0-10), exploring in particular the dimension of control, that is particularly important to drive complex adaptive responses to stress (25).

-A symptoms list was used to score somatic complaints (score 0-50)

The questionnaires were tested on Italian students and patients (Niddomi P. Doctoral Thesis, Dept. of Psychology, University of Urbino, Italy)

In addition, objective competence was judged from the grades obtained at the examination (on a 0-30 score, as customary in Italian Universities).

Study of autonomic regulation : in both control and stress days, all subjects were studied in resting conditions (lying down with 15° back support) and also during active orthostatism, i.e. a standardized condition leading to sympathetic predominance (10). After a 10-minute period, allowed for stabilization, a 10 minutes rest period was obtained, to be followed by a further period of 7 minutes recording during active standing.

Recorded variables: using standard Ag-AgCl electrodes and a piezoelectric (PVF2, 3M, London, Ontario, Canada) respiratory belt, both the ECG (CM5) and the respiratory signal were monitored with a two way radiotelemetry system (Marazza, Monza, Milano, Italy). The arterial pressure waveform was continuously estimated non invasively with a plethysmographic device (Finapres, Ohmeda, Englewood, Colorado, USA). Data were acquired on a PC, using an acquisition rate of 300 samples/channel/second.

Data analysis: from the ECG derived tachogram an autoregressive approach provides both absolute (i.e. [msec²]) and normalized units (i.e. [nu]) spectral powers of Low (LF) and High Frequency (HF) components as previously described (10,11). Spectral analysis was also performed on the systolic arterial pressure and the respiratory signals using a similar procedure. This latter assessed the main respiratory frequency, which was always higher than 0.15 Hz. From the simultaneous analysis of arterial pressure and RR interval variabilities a frequency domain index (a) can be derived, which is a measure of the overall gain of the arterial pressure-heart period relationship, and provides results similar to those obtained with the phenylephrine slope approach (26). In keeping with previous studies from our laboratory, employing non invasive estimates of arterial pressure, the following formula providing an average value (27) was employed: a = [(PRR / PSAP)LF½ + (PRR / PSAP)HF½] / 2

Statistics : data are presented as mean ±SEM. Statistical tests included two way ANOVA for repeated measures, followed by individual contrasts, and the paired t test, as appropriate (Sigmastat, Jandel). An a level £ 0.05 was considered significant. Correlation analysis was used to estimate the statistical link between selected parameters treating data from the two experimental days as independent observations.